JP2014115961A - Information processor, vehicle identification method and vehicle identification program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To achieve appropriate vehicle identification.SOLUTION: An information processor includes: number recognition means for extracting the back and forth number plates of a vehicle from a photographic image, and for recognizing respective vehicle information included in the extracted back and forth number plates; and vehicle identification means for comparing the respective vehicle information acquired by the number recognition means for each of corresponding characters or numbers, and for, when the compared characters or numbers are different, acquiring a relevance ratio between each character or number preliminarily stored in storage means and the different characters or numbers, and for identifying the vehicle by using the acquired relevance ratio.

Description

  The present invention relates to an information processing apparatus, a vehicle identification method, and a vehicle identification program.

  Conventionally, a license plate that recognizes the number of a license plate (vehicle number) from a vehicle image taken with a camera installed on a road or parking lot, etc., for the control of vehicles due to speeding violations or the investigation of stolen vehicles, etc. There is a recognition system. In addition, license plates (hereinafter referred to as “plates”) are attached to the front and rear surfaces of vehicles, but in the case of stolen vehicles, the front plates that can be easily replaced are replaced and forged (spoofed). There are many. As a mechanism for detecting this forgery, for example, there is a method of capturing and recognizing the front and rear surfaces of a vehicle and determining whether the recognition results match (for example, see Patent Document 1).

JP 2004-206441 A

  However, since the plate number (for example, letters and numbers in the plate) is recognized from the photographed image, the number may not be recognized normally depending on the environmental conditions for photographing. In the method as shown in Patent Document 1 described above, when the front and rear numbers do not match, it cannot be determined whether it is due to forgery of the plate or abnormalities such as misreading at the time of number recognition. Vehicle identification cannot be realized.

  In one aspect, the present invention aims to achieve proper vehicle identification.

  An information processing apparatus according to an aspect is obtained by number recognition means for extracting license plates before and after a vehicle from a photographed image and recognizing respective vehicle information included in the extracted front and rear license plates, and the number recognition means. Each vehicle information is compared for each corresponding predetermined character or number, and when the compared character or number is different, each character or number stored in advance in the storage means and each different character or number Vehicle recognizing means for identifying the vehicle using the obtained relevance ratio.

  Appropriate vehicle identification can be realized.

It is a figure which shows the function structural example of a vehicle identification system. It is a figure which shows the function structural example of information processing apparatus. It is a figure which shows an example of the hardware constitutions which can implement | achieve a vehicle identification process. It is a flowchart which shows an example of a number recognition process. It is a flowchart which shows an example of a vehicle identification process. It is a figure which shows an example of a license plate. It is a figure which shows an example of the data memorize | stored in a memory | storage means. It is a figure (the 1) which shows the comparative example of the back-and-front plate. It is a figure (the 2) which shows the comparative example of the back-and-front plate. It is a figure (the 3) which shows the comparative example of the back-and-front plate. It is a figure which shows the example of determination with respect to the difference of a precision.

  Hereinafter, embodiments will be described in detail with reference to the accompanying drawings.

<Functional configuration example of vehicle identification system>
FIG. 1 is a diagram illustrating a functional configuration example of a vehicle identification system. A vehicle identification system 10 illustrated in FIG. 1 includes a plurality of imaging devices 11-1 and 11-2 (hereinafter collectively referred to as “imaging device 11” as necessary) and an information processing device 12. The imaging device 11 and the information processing device 12 are connected in a state in which data can be transmitted and received by a communication network 13 typified by, for example, a Local Area Network (LAN) or the Internet.

  The imaging device 11 captures the vicinity of the front and the rear of the vehicle 20 in order to capture the plates attached to the front and the rear of the vehicle 20 passing through a road, a parking lot, and the like. In the example of FIG. 1, the front of the vehicle 20 is photographed by the imaging device 11-1, and the rear of the vehicle 20 is photographed by the imaging device 11-2. However, the present invention is not limited to this. For example, the front and rear may be photographed using one imaging device 11, and the front and rear may be photographed using three or more imaging devices.

  The vehicle identification system 10 includes, for example, a detection sensor that detects the passage of the vehicle, and the imaging device 11 may perform shooting at the timing when the detection signal from the detection sensor is received. You may take a picture. The imaging device 11 transmits the captured image to the information processing device 12 via the communication network 13.

  The imaging device 11 may be, for example, a digital camera, and an infrared camera or a surveillance camera that can be photographed at night can be used, but is not limited thereto.

  The information processing device 12 performs vehicle identification from the captured image from the imaging device 11. Specifically, the information processing device 12 acquires two front and rear plate images of each vehicle 20 from the captured image, and acquires number information (vehicle information) described on each plate by character recognition or the like. . The number information includes, for example, a transportation branch office, a classification number, kana characters, a series designation number, etc., but is not limited thereto.

  In addition, the information processing device 12 collates the front and back plates, and when the number information is different, for example, extracts the precision set in advance for each target character, and the difference between the top two values having the highest precision To get. Further, the information processing apparatus 12 determines, for example, whether the plate is forged or due to an abnormality such as misreading at the time of number recognition from the obtained difference. Thereby, in this embodiment, appropriate vehicle identification is realizable.

  Further, when it is determined that the number recognition process is not abnormal and the recognition results of the preceding and following plates are different, the information processing apparatus 12 determines the position and number of the different characters and numbers. Based on this, the type (level) of plate counterfeiting can be determined. The type of forgery is, for example, replacement of a plate or partial forgery of some characters or numbers, but is not limited thereto.

  In this embodiment, when the collation results of the numbers of the front and rear plates of the vehicle are different, the degree of plate counterfeiting and the number recognition are compared by comparing the relevance ratios of the recognition results of the number information described on the plates. Appropriate vehicle identification can be realized by presenting the degree of abnormality of the processing.

  The information processing apparatus 12 is, for example, a general-purpose personal computer (PC) or a server, but is not limited thereto. The communication network 13 may be wired or wireless, or a combination thereof.

<Example of Functional Configuration of Information Processing Device 12>
Next, a functional configuration example of the information processing apparatus 12 described above will be described with reference to the drawings. FIG. 2 is a diagram illustrating a functional configuration example of the information processing apparatus. The information processing apparatus 12 shown in FIG. 2 includes number recognition means 30, vehicle identification means 40, storage means 50, and display means 60. In the example of FIG. 2, imaging devices 11-1 to 11-n are also shown.

  The number recognition unit 30 includes an image acquisition unit 31, a vehicle extraction unit 32, a plate extraction unit 33, a number recognition unit 34, and a control unit 35. Further, the vehicle identification means 40 includes a plate recognition result acquisition means 41, a front / rear plate comparison means 42, a precision ratio difference calculation means 43, a recognition processing determination means 44, a determination result reflection means 45, and a plate abnormality rate calculation means. 46 and control means 47.

  The image acquisition unit 31 acquires front and rear images including the plate of the vehicle 20 from the one or more imaging devices 11-1 to 11-n via the communication network 13.

  The vehicle extraction unit 32 extracts a vehicle from the captured image obtained from the imaging device 11. Specifically, image processing such as edge detection is performed on the captured image, the shape of the object obtained by the image processing is compared with a preset vehicle shape, and the shapes match or exceed a predetermined threshold value If the degree of similarity is, an object having that shape is extracted as a vehicle.

  The vehicle extraction means 32 may repeat the image processing such as edge detection described above, for example, while changing the content of the image processing until the vehicle shape is extracted, or may be performed only a predetermined number of times. Changing the contents of image processing includes, for example, relaxing edge detection conditions to detect more edge parts, or conversely reducing conditions to reduce edge detection parts, but is not limited to this. Absent. For example, other image processing (for example, smoothing processing or sharpening processing) may be combined.

  Further, the vehicle extraction unit 32 extracts an object having the shape as a vehicle when the object shape matches or is similar to a preset vehicle shape without performing the image processing such as edge extraction described above. May be.

  In addition, when the vehicle area included in the image can be specified from the relationship between the installation position of the imaging device 11 and the passing position of the vehicle 20, the specified area may be extracted as the area of the vehicle 20. The method for extracting the vehicle 20 is not limited to this.

  The plate extraction unit 33 extracts a plate portion from the image area of the vehicle 20 extracted by the vehicle extraction unit 32. The plate portion has, for example, a rectangular shape. Therefore, the plate extraction means 33 determines whether or not a predetermined rectangular area exists in the vehicle area, and when the predetermined rectangular area exists, extracts the area as an image of the plate portion.

  Note that the size of the rectangle described above may vary depending on the position of the imaging device 11 and the vehicle 20, the shooting magnification, the vehicle type, and the like. Therefore, the plate extraction means 33 may estimate the rectangular area of the plate from the size of the vehicle relative to the vehicle area, for example, and extract the plate portion using the rectangular area. The rectangular area may have a predetermined allowable range, and the extraction process may be performed including the allowable range.

  Further, as a method other than the above, the plate extraction unit 33 may previously provide a plurality of rectangular region patterns having different sizes and extract the plate portion by matching with each region pattern.

  In the present embodiment, depending on the positional relationship between the imaging device 11 and the vehicle 20, the plate in the image may be photographed with an inclination. In such a case, the plate extraction means 33 inclines the rectangular shape in accordance with, for example, the inclination (angle) obtained from the shape of the vehicle 20 in the captured image, and extracts the plate portion from the inclined area. May be.

  The number recognizing unit 34 performs character recognition processing on the plate extraction areas before and after the vehicle obtained by the plate extracting unit 34, and acquires number information (vehicle information). The number recognition unit 34 stores the obtained number information in the storage unit 50. At this time, the number recognizing means 34 stores the number information obtained from the front and rear plates for the vehicle 20 in the storage means 50 in association with each other. The number recognition unit 34 may directly output the recognized number information to the vehicle identification unit 40.

  The control means 35 controls the entire components in the number recognition means 30. Specifically, the control means 35 performs each control regarding recognition of number information, for example based on the input of the picked-up image from the imaging device 11, the starting instruction by a user, etc. Here, each control refers to, for example, image acquisition by the image acquisition unit 31 described above, vehicle extraction by the vehicle extraction unit 32, plate portion extraction by the plate extraction unit 33, number information recognition by the number recognition unit 34, and the like. There is control, but it is not limited to this.

  Further, in the vehicle identification unit 40 shown in FIG. 2, the plate recognition result acquisition unit 41 acquires the plate number information recognized by the number recognition unit 30 described above and stored in the storage unit 50. In addition, the plate recognition result acquisition means 41 acquires the number information of the plate before and behind with respect to the same vehicle.

  The front / rear plate comparison means 42 compares the plate information before and after acquired by the plate recognition result acquisition means 41. In addition, the comparison of plate information is performed using at least one item among, for example, a transportation branch office, a classification number, a kana character, and a series designation number included in the number information. The front and rear plate comparing means 42 does not need to perform the precision ratio difference calculation means 43 or the like when the number information of a predetermined threshold value or more is different, and immediately outputs it to the recognition process determination means 44. It can be determined that the number has been forged.

  The relevance ratio difference calculation means 43 calculates the relevance ratio for different characters and numbers from the relevance ratio for each character and number stored in advance in the storage means 50 for each of different characters and numbers in the comparison result by the front and rear plate comparison means 42. Get high letters and numbers. The precision ratio difference calculation means 43 calculates the precision ratio difference set for characters and numbers having a high precision ratio.

  Here, the relevance ratio represents, for example, the degree of similarity with a predetermined font (template) that serves as a model for each piece of reference number information (letters and numbers). The relevance ratio can be acquired in advance by template matching, which is a general image processing technique, and the acquired similarity value is stored in the storage unit 50. In this embodiment, all the templates of each number information (for example, 0 to 9 in the case of numbers) are obtained, all the matching rates with each template are obtained, and the one with the highest matching rate is set as the recognition result. In the case of having a plurality of fonts, the matching rate is obtained for the template.

  The recognition processing determination unit 44 determines whether the front and rear plates are matched or forged based on the comparison result by the front and rear plate comparison unit 42. Further, the recognition processing determination unit 44 obtains the number (ratio) of characters and numbers having the same (or different) number information with respect to the total number of characters (including numbers) to be compared, for example, and the whole plate is different. It is also possible to determine whether or not a part is different.

  In addition, the recognition processing determination unit 44 determines that the number recognition processing is not abnormal and determines that the recognition results of the preceding and following plates are different, the position and number of the different characters and numbers are changed. Based on this, the type (level) of plate counterfeiting can be determined. The type of forgery is, for example, replacement of a plate or partial forgery of some characters or numbers, but is not limited thereto.

  Further, the recognition process determination unit 44 determines whether there is an abnormality in the recognition process based on, for example, the result of the relevance ratio difference by the relevance ratio difference calculation unit 43. For example, when there is an abnormality due to misreading of characters in the recognition process, the recognition process determination unit 44 causes the plate abnormality rate calculation unit 46 to calculate the abnormality rate.

  Furthermore, the recognition process determining unit 44 performs the process in the above-described relevance ratio difference calculating unit 43 when the degree of the recognition process abnormality obtained by the plate abnormality rate calculating unit 46 is higher than a preset threshold value. May be excluded. Note that these various controls are performed by the control means 47.

The determination result reflection unit 45 stores the determination result obtained by the recognition process determination unit 44 in the storage unit 50. That is, the determination result reflecting unit 45 updates the matching rate of each character or number stored in the storage unit 50 based on the newly obtained authentication result with respect to the current matching rate. 46 is a plate as an index of how abnormal the plate is when the recognition process determining unit 44 determines that the plate is counterfeited by the difference between the values of the matching rate obtained by the matching rate difference calculating unit 43. Calculate the abnormal rate. Specifically, the plate abnormality rate calculating means 46 calculates, for example, the plate abnormality rate (the degree that the plate is counterfeit) from the ratio of different characters and numbers to the total number information obtained from the comparison results. However, the calculation method is not limited to this.

  The control unit 47 controls the entire components in the vehicle identification unit 40. Specifically, the control means 47 performs each control regarding vehicle identification based on the instruction | indication by a user etc., for example. Here, each control refers to, for example, acquiring the plate recognition result by the plate recognition result acquisition means 41 described above, comparing the front and rear plates by the front and rear plate comparison means 42, and the precision difference difference calculation means 43 by the precision ratio difference calculation means 43. For example, to calculate. In addition, each control refers to, for example, determining whether the front and rear plates are coincident by the recognition processing determination unit 44, reflecting the determination result in the storage unit 50 by the determination result reflection unit 45, or by the plate abnormality rate calculation unit 46. There are controls such as calculating the plate abnormality rate.

  Moreover, the control means 47 is based on the result obtained from each component part in the vehicle identification means 40, and when the front and rear number recognition results do not match, it is due to forgery of the plate or abnormalities such as misreading at the time of number recognition. A vehicle identification result such as an object is output to the storage means 50. Further, when the number information of the front and rear plates all match, the control means 47 determines that the vehicle to which the plates are attached is a normal vehicle and outputs the vehicle identification result to the storage means 50. The vehicle identification result described above may be output to the display means 60. In addition, the control content in the control means 47 is not limited to this.

  The storage unit 50 stores various types of information required in the present embodiment. Specifically, the storage unit 50 stores, for example, images acquired from the imaging device 11, processing progress and processing results by the number recognition unit 30, processing progress and processing results by the vehicle identification unit 40, and the like.

  Further, the storage unit 50 stores information for extracting the vehicle by the vehicle extraction unit 32, information for extracting the plate by the plate extraction unit 33, and the like. In addition, the storage unit 50 stores a precision rate for each character or each number set in advance. The information stored in the storage unit 50 is not limited to the information described above.

  The storage unit 50 can read and write various stored information at a predetermined timing as necessary. The storage unit 50 is a collection of various types of information as described above, and is a database systematically configured so that the information can be searched and extracted using, for example, keywords. You may have the function of. The storage means 50 is composed of, for example, a hard disk or a memory.

  The display means 60 displays various information obtained by this embodiment on a screen or the like. For example, the display unit 60 displays an image or video obtained by the photographing apparatus 11, a number recognition result, a calculation result of a matching rate or a plate abnormality rate, a determination result, or the like. The determination result is, for example, whether the number information of the front and rear plates does not match, is forgery of the plate, is due to an abnormality such as misreading at the time of number recognition, or is an appropriate vehicle (for example, It is a result of whether the vehicle is a stolen vehicle, but is not limited to this. The display means 60 is composed of, for example, a display or a monitor.

  As described above, according to the information processing apparatus 12, when the recognition results of the number information by the plurality of plate images are different, each number is actually different by comparing each matching rate of the corresponding letters and numbers. The possibility of being present can be determined. Further, the information processing apparatus 12 can determine the possibility of an abnormality due to misreading of the recognition process or the like by comparing the relevance ratios described above. In addition, the information processing apparatus 12 can calculate a ratio of matching characters and numbers in the recognition result and determine whether a part of the characters is different or the whole is different.

  In other words, according to the present embodiment, when the recognition results of a plurality of images (for example, preceding and following license plate images) do not match, one of the targets of the images has an abnormality (for example, forgery such as replacement) as an index. Can be presented as a value. Further, when the recognition results of a plurality of images do not match, the possibility that the recognition result is due to an abnormality in the number recognition process can be presented as an index value. In addition, it is possible to determine whether the abnormality of one of the objects of the image is due to forgery such as replacement or due to partial forgery of characters or the like.

<Hardware Configuration Example of Information Processing Device 12>
Here, in the information processing apparatus 12 described above, an execution program (vehicle identification program) that can cause a computer to execute each function is generated, and the execution program is installed in, for example, a general-purpose PC or server. The vehicle identification process in this embodiment can be realized. Here, a hardware configuration example of a computer capable of realizing the vehicle identification process in the present embodiment will be described with reference to the drawings.

  FIG. 3 is a diagram illustrating an example of a hardware configuration capable of realizing the vehicle identification process. 2 includes an input device 71, an output device 72, a drive device 73, an auxiliary storage device 74, a main storage device 75, a central processing unit (CPU) 76 that performs various controls, and a network connection. Device 77 and these are connected to each other by a system bus B.

  The input device 71 has a pointing device such as a keyboard and a mouse operated by a user or the like, and a voice input device such as a microphone, and activates a program execution instruction, various operation information, software, or the like from the user or the like. Enter the information.

  The output device 72 has a display for displaying various windows, data, and the like necessary for operating the computer main body for performing the processing in the present embodiment. Can be displayed. Further, the output device 72 can print the above processing result or the like on a print medium such as paper and present it to the user or the like. A part of the output device 72 corresponds to, for example, the display unit 60 described above.

  Here, the execution program installed in the computer main body in the present embodiment is provided by, for example, a universal serial bus (USB) memory, a portable recording medium 78 such as a CD-ROM, a DVD, or the like. The recording medium 78 on which the program is recorded can be set in the drive device 73, and an execution program included in the recording medium 78 is transferred from the recording medium 78 to the auxiliary storage device 74 via the drive device 73 based on a control signal from the CPU 76. To be installed.

  The auxiliary storage device 74 is a storage means such as a hard disk, and stores the execution program in the present embodiment, a control program provided in the computer, etc. based on a control signal from the CPU 76, and performs input / output as necessary. be able to. The auxiliary storage device 74 can read and write necessary information from each stored information based on a control signal from the CPU 76 and the like.

  The main storage device 75 stores an execution program or the like read from the auxiliary storage device 74 by the CPU 76. The main storage device 75 includes a read only memory (ROM), a random access memory (RAM), and the like. A part of the auxiliary storage device 74 and the main storage device 75 corresponds to the storage means 50 described above, for example.

  The CPU 76 controls processing of the entire computer, such as various calculations and data input / output with each hardware component, based on a control program such as an operating system and an execution program stored in the main storage device 75. Each process can be realized. Various information necessary during the execution of the program can be acquired from the auxiliary storage device 74, and the execution result and the like can also be stored.

  Specifically, the CPU 76 corresponds to the program on the main storage device 75 by causing the vehicle identification program installed in the auxiliary storage device 74 to be executed based on, for example, a program execution instruction obtained from the input device 71. Process. For example, CPU76 performs the process in the number recognition means 30 mentioned above, the process in the vehicle identification means 40, etc. by running a vehicle identification program. In addition, the processing content in CPU76 is not limited to this. The contents executed by the CPU 76 can be stored in the auxiliary storage device 74 as necessary.

  The network connection device 77 acquires an execution program, software, setting information, and the like from an external device or the like connected to the communication network 14 by connecting to the communication network 14 or the like based on a control signal from the CPU 76. In addition, the network connection device 77 can provide the execution result obtained by executing the program or the execution program itself in the present embodiment to an external device or the like.

  With the hardware configuration described above, the vehicle identification process in the present embodiment can be executed. Further, by installing the program, the vehicle identification process in the present embodiment can be easily realized by a general-purpose PC or server.

<About vehicle identification processing>
Here, an example of the vehicle identification process in this embodiment is demonstrated using a flowchart. In addition, since the vehicle identification process in this embodiment is divided roughly into a number recognition process and a vehicle identification process, each is demonstrated separately.

<Flowchart of number recognition process>
FIG. 4 is a flowchart illustrating an example of the number recognition process. In the example of FIG. 4, in the number recognition process, an image obtained by photographing a vehicle is acquired (S01), and a vehicle portion is extracted from the acquired image (S02). In the process of S02, for example, image processing such as edge detection is performed on the image obtained in the process of S01, and the shape of the object obtained from the result is compared with a preset shape of the vehicle. However, the present invention is not limited to this.

  Next, the number recognition process determines whether or not the vehicle has been successfully extracted (S03). If the vehicle has not been successfully extracted (NO in S03), the process returns to S01 to acquire another image. Vehicle extraction. In addition, at the time of vehicle extraction, it is not necessary to acquire a new image. For example, a vehicle may be extracted by performing image processing different from the previous image processing.

  Further, in the number recognition process, when the vehicle is successfully extracted (YES in S03), the plate included in the vehicle area is extracted (S04), and the number in the extracted plate area is recognized (S05). ). In the number recognition process, the number information recognition result obtained by the process of S05 is stored in the storage means 50 or the like (S06).

<Flowchart of vehicle identification process>
FIG. 5 is a flowchart illustrating an example of the vehicle identification process. In the example of FIG. 5, in the vehicle identification process, for example, the plate recognition result before and after the vehicle obtained by the number recognition process described above is input from the storage means 50 or the like (S11).

  Next, the vehicle identification process compares the front and rear plates (S12). The comparison of the front and rear plates is, for example, a comparison for each character and each number included in the number information of the front and rear plates. Further, in the vehicle identification process, it is determined whether or not the front and rear plates all match (S13). That is, in the process of S13, it is determined whether or not all the characters or numbers included in the number information of the preceding and following plates match.

  Here, the number information includes, but is not limited to, for example, at least one of the transportation branch, classification number, kana characters, series designation number, etc., for example, alphabetic characters instead of kana characters, etc. But you can. In the example of FIG. 5, it is determined whether the transportation branch office, the classification number, the kana character, and the series designation number all match.

  In the vehicle identification process, if all the front and rear plates do not match (NO in S13), a difference in the precision is calculated (S14), and whether or not the recognition process is abnormal based on the calculated precision is determined. A determination is made (S15).

  The vehicle identification process determines whether the degree of abnormality in the recognition process is equal to or greater than a preset threshold (S16). If the degree of abnormality is not equal to or greater than the threshold (NO in S16), for example, a plate Is determined to be due to counterfeiting, and the abnormal rate of the plate is calculated (S17).

  In the vehicle identification process, if the degree of abnormality is equal to or greater than the threshold (YES in S16), the vehicle identification process is determined as an abnormality candidate due to misreading of the number recognition process (S18). In the process of S18, for example, the degree of abnormal recognition may be calculated as one of the indices. The degree to which the recognition is abnormal becomes higher as the difference in the relevance ratio is smaller. Therefore, for example, by calculating as “degree = 100−difference (%) of relevance ratio”, the degree of abnormal recognition can be obtained, but the calculation method is not limited to this.

  Further, the vehicle identification process determines that the vehicle is a normal vehicle if all the front and rear plates match in the process of S13 described above (YES in S13) (S19).

  In the vehicle identification process, the vehicle identification result is stored in the storage unit 50 or the like after any of the above-described processes of S17 to S19 is completed (S20). By the process of S20, the precision of each character stored in the storage unit 50 is updated, and the precision is reflected in the next vehicle identification process. The vehicle identification process may include a process of displaying the vehicle identification result obtained in S17 to S19 on a screen such as a display unit.

<Example of vehicle extraction>
Here, an example of the vehicle extraction described above will be described. The vehicle extraction unit 32 described above acquires a captured image captured from the imaging device 11 and performs edge detection on the acquired captured image.

  For example, as edge detection, a location where the brightness of a digital image changes sharply (discontinuously) is specified. By performing edge detection in this way, it is possible to greatly reduce the amount of data to be processed, and it is possible to extract only structural attributes of an image while eliminating relatively less important information.

  Thereby, in this embodiment, by comparing the shape obtained by the edge detection with the preset vehicle shape stored in the storage unit 50 or the like, the shapes match or are similar with a similarity equal to or greater than a predetermined threshold. The existing shape can be extracted as the vehicle portion. In the present embodiment, the present invention is not limited to this, and the vehicle may be extracted by other methods, or the vehicle may not be extracted.

<About the number plate configuration>
Next, an example of number information recognized by the number recognition unit 34 will be described with reference to the drawings. FIG. 6 is a diagram illustrating an example of a license plate. The number plate 80 shown in FIG. 6 includes a transportation branch office 81, a classification number 82, a kana character 83, and a series designation number 84 as number information.

  The license plate 80 is a sign plate attached to identify individual vehicles. In addition, the size of the license plate 80 varies depending on the type and size of the vehicle (for example, a large scale plate, a medium size plate, a small size plate, etc.). Therefore, in the present embodiment, for example, the plate area of the rectangular area can be searched and extracted in accordance with the size of the vehicle acquired at the time of vehicle extraction, the vehicle type, and the like. In addition, the plate extraction method is not limited to this, for example, by setting a template of a rectangular area corresponding to a preset plate size and performing matching with each template, A plate region may be extracted.

  In addition, the color scheme of the plate varies depending on “automobile private use”, “automobile business use”, “light vehicle private use”, “light vehicle business use”, etc., but in this embodiment, the plate area is extracted regardless of the color arrangement. In addition, color information may be extracted during plate extraction.

  Note that the information of the license plate described above is stored in the storage means 50. In this embodiment, each of the transportation branch 81, the classification number 82, the kana character 83, and the series designation number 84 is analyzed by character recognition or the like. Based on the analysis result, the vehicle identification means 40 determines the number information of the front and rear plates. Compare these. In addition, although it is preferable to compare the classification number 82 and the series designation number 84 for every number character, it is not limited to this. Moreover, although it is preferable that the transportation branch 81 is compared for every place name and the kana character 83 is compared for every character, it is not limited to this. The kana character 83 may be an alphabetic character or the like.

  Further, in the present embodiment, in the number recognition process, in addition to the numbers constituting the plate, the plate size (large size, medium size, etc.), color (for private use, for business use, etc.), and the light type plate are used. It is also possible to recognize the characteristics of the plate such as whether or not, and the characteristics of the vehicle such as the shape of the vehicle body (truck, wagon, sedan, etc.) and the color being painted (black, white, blue, etc.).

  In the present embodiment, when comparing the number recognition results, it may be compared whether the features (size, color, etc.) of the plates match between the front and rear plates. Thereby, by adding the feature of the plate to the comparison item, it is possible to improve the determination accuracy such as whether the entire plate is replaced.

  Furthermore, in this embodiment, it can be determined whether the front and rear plates of the same vehicle are reliably compared by confirming that the characteristics of the vehicle 20 match in front and rear.

<Data configuration example of storage means 50>
Next, a data configuration example of the storage unit 50 will be described with reference to the drawings. FIG. 7 is a diagram illustrating an example of data stored in the storage unit. In the example of FIG. 7, the number information data recognized by the number recognition process, the presence / absence of a preset precision corresponding to each character and each number of the data, and an example of the precision are shown. Yes.

  The configuration example of the data shown in FIG. 7 includes “authentication ID”, “recognition date / time”, “number information”, etc. as items, but the type, number, order, format, etc. of the items are It is not limited. In addition, as the number information, etc., “Transportation Branch etc.”, “Category Number 1st Digit”, “Category Number 2nd Digit”, “Category No. 3rd Digit”, “Kana Character”, “Sequence Specified Number 1st Digit” ”,“ Second digit of serial designation number ”,“ third digit of serial designation number ”,“ fourth digit of serial designation number ”, etc., but are not limited thereto. Further, in the storage means 50, the content of “data” is set for the item, and “adaptation rate” is set for the number information.

  In the example of FIG. 7, the ID for identifying each recognized plate is “1100011”, and the recognized date / time is “2012/07/01 00:12:34”. In addition, it is shown that the number information “Transportation Branch etc.” is “Kawasaki” and the “Relevance Rate” is “Yes”. In the example of FIG. 7, the first digit of the classification number is “5”, the precision is “Yes”, the second digit of the classification number is “5”, the precision is “Yes”, and the third digit of the classification number is “−”. "The precision is" Yes ", the kana character is" F "and the precision is" Yes ". In the example of FIG. 7, the first digit of the series designation number is “·” and the precision is “Yes”, the second digit of the series designation number is “·” and the precision is “Yes”, and the third digit of the series designation number. Is “2”, the precision is “Yes”, the fourth digit of the series designation number is “2”, and the precision is “Yes”.

  Here, “existence” of the matching rate means having a matching rate that has been processed using a predetermined template in advance. That is, as shown in FIG. 7, for example, when the matching rate of the fourth digit in the series designation number is “present”, in the sub-table, the fourth digit number and the matching for each number included in the preset template The rate is displayed. In the present embodiment, the degree of similarity of each number of a predetermined font with respect to the number “2” in the fourth digit of the series designation number is displayed as the matching rate. By using this precision, for example, the first letter and precision (in the example of FIG. 7, "2" precision 92%) in the descending order of precision, and different letters and numbers in the preceding and following plate comparisons. Corresponding precision can be extracted.

  7 may be set for each plate or for a plurality of preset fonts, for example. When it is set for each font, for example, the highest matching rate (first matching rate) or the like can be obtained from each matching rate of a plurality of fonts, but the present invention is not limited to this. In addition, the sub-table of the relevance ratio may be changed to another sub-table depending on, for example, each vehicle type and every hour (for example, daytime, nighttime).

<Comparison example of front and rear plates>
Next, an example of front and rear plate comparison in the front and rear plate comparison means 42 will be described with reference to the drawings. FIGS. 8-10 is a figure (the 1-the 3) which shows the comparative example of a back-and-front plate. 8 to 10 are processes performed by the vehicle identification unit 40 described above, for example.

  FIG. 8 shows a case where a part of the plate is changed. For example, comparing the rear plate shown in FIG. 8 (A) with the front plate shown in FIG. 8 (B), the first and second digits of “..” and “11” in the series designation numbers are different. Yes. In this case, for the different values “·” and “1”, the respective matching ratios of the first and second digits of the serial designation number are acquired.

  In this case, in the example of the precision shown in FIG. 8C, in the case of the rear plate shown in FIG. 8A, the precision of “•” (“•” and “•” is used in the first digit of the serial number. 96) and the relevance ratio of “1” (similarity between “•” and “1”) is 24%. Therefore, the difference between the two precisions is 96−24 = 72%. In the second digit of the serial number, the precision of “•” is 96% and the precision of “1” is 24%. Therefore, the difference between the two precisions is 96−24 = 72%.

  On the other hand, in the case of the front plate shown in FIG. 8B, the precision of “·” is 12% and the precision of “1” is 92% in the first digit of the serial number. Therefore, the difference between the two precisions is 92-12 = 80%. Further, in the second digit of the serial number, when the matching rate of “•” is 21% and the matching rate of “1” is 94%, the difference of the matching rates is 94-21 = 73%.

  Here, FIG. 11 is a diagram illustrating a determination example with respect to the difference of the relevance ratio. Here, as a threshold value, for example, if a difference of 50% or more of the relevance ratio is “difference / large”, the difference of the relevance ratio is large in both FIG. 8A and FIG. Therefore, it can be determined from the example of FIG. 11 that the possibility of forgery is high. This determination can also indicate that the recognition of the number information is normal in both FIG. 8 (A) and FIG. 8 (B).

  In the example of FIG. 8, the values are different at two locations (the first digit and the second digit of the serial designation number). In that case, for example, when a preset number of items (for example, 2) or less is different, a result such as “partially different” can be output.

  In the case of a stolen car or the like, there are many cases in which the serial number that is large and conspicuous is changed. Therefore, for example, the vehicle identification result may be changed and output depending on whether the value is different depending on the serial number or whether the value is different from the serial number.

  For example, if the serial numbers are different by a predetermined number (for example, 2) or less, “partially different” may be set and a message indicating that the possibility of forgery is high may be output. Further, when the information other than the serial number is different by a predetermined number (for example, 2) or less, the vehicle identification result may be output as, for example, “there is a possibility of recognition abnormality”.

  Next, the example of FIG. 9 will be described. The number information of the rear plate shown in FIG. 9 (A) is “Kawasaki 55..22”, and the number information of the front plate shown in FIG. 9 (B) is “Shinagawa 500 8888”. As shown in FIGS. 9A and 9B, when almost all pieces of number information are different, it is not necessary to perform processing such as a difference in precision, and it is immediately determined that the number has been forged. be able to. In the process of FIG. 9, for example, it is determined that the number is immediately counterfeited based on, for example, a case where different portions of the number information differ by 80% or more of the entire number information or a case where the number information differs by a predetermined number or more. May be.

  That is, in the example of FIG. 9, for example, when the result of a predetermined number of items (for example, 4) or more among the items of the number information shown in FIG. Is output ”.

  Next, the example of FIG. 10 will be described. As shown in FIG. 10A and FIG. 10B, the result may be different due to an abnormality caused by misreading in the recognition process. For example, in FIG. 10A and FIG. 10B, the kana characters “fu” and “ro” are different. In this case, each difference is calculated from the relevance ratio of the kana characters included in the storage unit 50.

  In this case, in the example of the precision shown in FIG. 10C, in the case of the rear plate shown in FIG. 10A, in the kana characters, the precision of “F” is 81% and the precision of “RO” is 69%. Therefore, the difference between the two matching rates is 81−69 = 12%. On the other hand, in the case of the front plate shown in FIG. 10B, in the kana characters, the matching rate of “F” is 76% and the matching rate of “RO” is 85%. Therefore, the difference (absolute value difference) between the two matching rates is | 76−85 | = 9%. That is, the difference between FIG. 10A and FIG. 10B is smaller than the predetermined threshold (50%). Therefore, it can be seen from the determination criteria shown in FIG. 11 that there is a high possibility of recognition abnormality. In the example of FIG. 10, there is a possibility of forgery. Further, this determination can simultaneously indicate that the possibility of recognition abnormality is high in both FIG. 10 (A) and FIG. 10 (B).

  That is, in the example of FIG. 10, for example, when the difference (or an absolute value) between the relevance ratios of the characters in the preceding and following recognition results is a predetermined threshold (for example, 50%) or less, Together with “There is a difference”, it can be output as “There is a possibility of recognition abnormality”.

  This corresponds to a case where a recognition abnormality is a candidate. In this case, a user (administrator or the like) may visually check and reflect whether the recognition abnormality is present in the data.

<About other examples of vehicle identification processing using precision ratio>
Here, another example of the vehicle identification process using the relevance ratio in the present embodiment will be described. For example, as shown in the above-described example, when the precision ratio that is the first place of different values (characters and numbers) is acquired, if the value itself is small, the effect of dirt or shadows, or forgery accuracy (font Etc.). Therefore, if the value of the relevance ratio that is the first place is equal to or less than a preset threshold value, the vehicle identification result may be output with the number “unknown”.

  In addition, in the normal recognition result with a normal plate, the first matching rate is high, and the difference from the second matching rate is also large. Moreover, when the difference of the relevance ratio between the first place and the second place is small, there is a high possibility of recognition abnormality. Therefore, in the present embodiment, for example, the first and second relevance ratios are acquired, and when the difference is equal to or less than a preset threshold value, “unknown” or “first candidate, second candidate” Both results may be output.

  In other words, in the present embodiment, for example, the first and second conformances from the higher relevance ratio to the recognition result of the number information before and after (for example, “F” and “RO” in FIG. 10). The rate may be acquired, and the vehicle identification may be determined for the difference of the relevance rate.

  Also, for example, when the shapes of characters and numbers are similar, such as “ki” and “sa”, “wa” and “ne”, “2” and “7”, etc., the difference between the relevance ratios is small. Tend. For this reason, in the present embodiment, comparison of number information may be performed by setting a threshold value of a difference serving as a reference for determination variably according to the similarity of characters and numbers.

  As described above, according to the present embodiment, appropriate vehicle identification can be realized. Specifically, when the recognition results of a plurality of images (for example, plate images before and after the vehicle) do not match, the possibility that there is an abnormality in one of the targets of the images can be presented as an index value. Further, in the present embodiment, when the recognition results of the preceding and following plate images do not match, the possibility that the number recognition process is abnormal can be presented as an index value. Further, in the present embodiment, it is possible to exclude those having different recognition results from those having different recognition results. Further, in the present embodiment, it can be determined whether the abnormality of the object of the image is due to forgery such as replacement or due to partial forgery of characters and the like.

  Each embodiment has been described in detail above. However, the present invention is not limited to the specific embodiment, and various modifications and changes other than the above-described modification are possible within the scope described in the claims. .

In addition, the following additional remarks are disclosed regarding the above Example.
(Appendix 1)
Number recognition means for extracting the number plates before and after the vehicle from the photographed image and recognizing each vehicle information included in the extracted number plates before and after the vehicle,
Each vehicle information obtained by the number recognition means is compared for each corresponding predetermined character or number, and when the compared character or number is different, it differs from each character or number stored in the storage means in advance. An information processing apparatus comprising: a vehicle identification unit that acquires a matching rate of each character or number and identifies the vehicle using the acquired matching rate.
(Appendix 2)
The vehicle identification means includes
The information processing apparatus according to appendix 1, wherein a difference value of the relevance ratio corresponding to the different characters or numbers is calculated, and the vehicle is identified using the calculated difference value.
(Appendix 3)
The vehicle identification means includes
The information processing apparatus according to claim 2, further comprising: an abnormality rate calculating unit that calculates a degree of abnormality when the vehicle information is recognized by the number recognition unit from the difference value.
(Appendix 4)
The vehicle identification means includes
The information processing apparatus according to any one of appendices 1 to 3, wherein the information processing apparatus identifies whether the license plate is counterfeited or an abnormality in recognition performed by the number recognition unit.
(Appendix 5)
The vehicle identification means includes
Supplementary note 1 wherein the vehicle is identified without acquiring the relevance ratio when the number of different characters or numbers is equal to or greater than a predetermined threshold in the comparison of the vehicle information. 5. The information processing apparatus according to any one of items 4 to 4.
(Appendix 6)
The vehicle identification means includes
The number of forgery of the license plate is determined based on the position or number of different characters or numbers in the vehicle information of the front and rear license plates. Information processing device.
(Appendix 7)
The information processing apparatus according to any one of appendices 1 to 6, wherein the vehicle information is at least one of a transportation branch, a classification number, a kana character, and a series designation number.
(Appendix 8)
A number recognition step of extracting license plates before and after the vehicle from the captured image and recognizing each vehicle information included in the extracted front and rear license plates;
Each vehicle information obtained by the number recognition step is compared for each corresponding predetermined character or number, and when the compared character or number is different, it differs from each character or number stored in the storage means in advance. And a vehicle identification step of identifying the vehicle using the obtained precision rate.
(Appendix 9)
Extract the license plates before and after the vehicle from the captured image, recognize the vehicle information contained in the extracted license plates before and after,
Each vehicle information obtained by the number recognition step is compared for each corresponding predetermined character or number, and when the compared character or number is different, it differs from each character or number stored in the storage means in advance. A vehicle identification program for causing a computer to execute a process of acquiring a matching rate of each letter or number and identifying the vehicle using the acquired matching rate.

DESCRIPTION OF SYMBOLS 10 Vehicle identification system 11 Imaging apparatus 12 Information processing apparatus 13 Communication network 20 Vehicle 30 Number recognition means 31 Image acquisition means 32 Vehicle extraction means 33 Plate extraction means 34 Number recognition means 35 Control means 40 Vehicle identification means 41 Plate recognition result acquisition means 42 Front / rear plate comparison means 43 Precision ratio difference calculation means 44 Recognition process determination means 45 Determination result reflection means 46 Plate abnormality rate calculation means 47 Control means 50 Storage means 60 Display means 71 Input device 72 Output device 73 Drive device 74 Auxiliary storage device 75 Main Storage device 76 CPU
77 Network connection device 78 Recording medium 80 Number plate 81 Transportation branch office 82 Classification number 83 Kana character 84 Serial designation number

Claims (6)

Number recognition means for extracting the number plates before and after the vehicle from the photographed image and recognizing each vehicle information included in the extracted number plates before and after the vehicle,
Each vehicle information obtained by the number recognition means is compared for each corresponding predetermined character or number, and when the compared character or number is different, it differs from each character or number stored in the storage means in advance. An information processing apparatus comprising: a vehicle identification unit that acquires a matching rate of each character or number and identifies the vehicle using the acquired matching rate. The vehicle identification means includes
The information processing apparatus according to claim 1, wherein a difference value of the relevance ratio corresponding to the different characters or numbers is calculated, and the vehicle is identified using the calculated difference value. The vehicle identification means includes
3. The information processing apparatus according to claim 1, wherein the information processing apparatus identifies whether the license plate is forged or abnormal in recognition by the number recognition unit. 4. The vehicle identification means includes
The vehicle is identified without acquiring the relevance ratio when the number of different characters or numbers is equal to or greater than a predetermined threshold in the comparison of the respective vehicle information. The information processing apparatus according to any one of 1 to 3. A number recognition step of extracting license plates before and after the vehicle from the captured image and recognizing each vehicle information included in the extracted front and rear license plates;
Each vehicle information obtained by the number recognition step is compared for each corresponding predetermined character or number, and when the compared character or number is different, it differs from each character or number stored in the storage means in advance. And a vehicle identification step of identifying the vehicle using the obtained precision rate. Extract the license plates before and after the vehicle from the captured image, recognize the vehicle information contained in the extracted license plates before and after,
Each vehicle information obtained by the number recognition step is compared for each corresponding predetermined character or number, and when the compared character or number is different, it differs from each character or number stored in the storage means in advance. A vehicle identification program for causing a computer to execute a process of acquiring a matching rate of each letter or number and identifying the vehicle using the acquired matching rate.

Images